Use a modifier with find command's regex?

Question

I'm trying to use the find command's regex to match newlines using the dot character. This usually is accomplished via the s modifier; however, different regex engines have different formats and I can't figure out exactly how to write the regex so that the find command will match newlines. For example, how would I change the following simple command to match all files including those containing newlines in their names:

find /home/user -regex ".*"

I've tried "(?s).*", "/.*/s", and probably a few others.

Answer 1

My version of GNU find supports a few different regex variants, which can be chosen with -regextype:

valid types are ‘findutils-default’, ‘awk’, ‘egrep’, ‘ed’, ‘emacs’, ‘gnu-awk’, ‘grep’, ‘posix-awk’, ‘posix-basic’, ‘posix-egrep’, ‘posix-extended’, ‘posix-minimal-basic’, ‘sed’.

Almost all of those let the dot match a newline, ...except emacs, which is the default.

So, use some other one. posix-extended is probably sane enough:

$ touch hello $'new\nline'
$ find . -type f -regex '.*'
./hello
$ find . -type f -regextype posix-extended -regex '.*'
./hello
./new?line

If you're running some other find, your mileage may vary. On a quick test, the one on my Mac does match newlines with . by default.

Then again, for simple patterns, you could just use -name which takes shell globs (and shouldn't have issues with newlines).

The (?s) looks like a feature of PCRE/Perl regexes, and I don't see find supporting that. As for /.*/s, find takes just the raw pattern, no separators, and hence no trailing options.

Answer 2

The command

find ~user

would find all names, including those containing newlines. Add -type f if you want to only find regular files.

If you don't want to restrict the matching of names, don't use a test on the names.

To restrict the names matched using substrings, you could use a globbing pattern with the -name test. Both * and ? matches newlines in a globbing pattern, both in the shell in general and in the -name test of find:

Example using standard find:

$ find . -name 'file?name'
./file
name

$ find . -name 'file*name'
./file
name

$ find . -name 'd*name'
./dir
name

$ echo ./d*name
./dir
name

---

Answer to misunderstood question ("How may I find names containing newlines"):

Using standard find:

find ~user -name '*
*'

Using "C-strings" in shells that support these (still using standard find)

find ~user -name $'*\n*'

None of the built-in regular expression types normally available in GNU find is able to match a newline with \n, so you're left using a literal newline, just like in the globbing patterns above. This means it's probably easier to use a standard -name test than a non-portable -regex test.